Fuzzy Neural Network Image Weak Target Detection

The detection and tracking of moving dim small target is highly focused in the field of computer vision and image processing. It also plays more and more important role in the system of infrared detection and surveillance, especially in the condition of complicated background and far distance. The main challenge of detection and tracking of dim point target comes from application of long-range surveillance system, such as wide FOV (field of view) telescope and SBIRS (Space-based infrared system) and other surveillance system. The dim small target usually hides in the constructed background clutter and heavy noise environment. In recent years, researchers pay much attention on detection and tracking of dim target in IR image sequence. Performance of algorithm is vital to the effective distance and intelligent extent of the IR surveillance system.The purpose of the paper is, in the first, the system is proposed, which is detecting dim small moving targets in image sequences based on background clutter suppression distinctly. This is a two-level policy, which is adopted that suppressing image clutter firstly and then detecting targets with matched filter. And the tasks of clutter suppression are summarized as two points: one is removing background energy and improving SCR, the other is transforming the complex, unknown, spatial-temporal non-stationary distribution of original image to simple, known one.An effective method is proposed to detecting dim small moving targets in infrared images, based on to Dynamic Fuzzy Neural Networks (D-FNN). The target of interest is assumed to have a very small spatial spread, and is obscured by heavy background clutter. The clutter is predicted and threshold exactly by D-FNN to and subtracted from the input signal, leaving components of the target signal in the residual. For the purpose of tracking complex background including different sub-structure, the raw image data is subdivided to some sub-block, in which the training samples are chosen for optimizing the weights of structuring element in the corresponding block, which leads to better clutter estimation Develop the target tracking method in the single spectrum dim moving target detection and tracking system which combined high speed scanning detector with staring detector by using common threshold technique, depending on the continuity and orientation of the target motivation, analyzing imaging characteristic of scanning detector and staring detector, using least mean square prediction technique.Analyze the detection probability and false alarm probability theoretically and buildup models and make experiments by using target detection algorithm performance evaluation method and corresponding precondition and hypothesis, considering target/background statistic characteristic.